stock-weekly-report 0.2.2 → 0.2.4

package/fetch_episodes.py

@@ -187,11 +187,6 @@ def fetch_and_download(config: dict, folder_name: str | None = None) -> None:
             filename = f"{program_name}_{date_str}{ext}"
             dest = speaker_dir / filename
 
-            if dest.exists():
-                print(f"  SKIP (already downloaded): {filename}")
-                found += 1
-                continue
-
             print(f"  Downloading: {filename}")
             try:
                 download_file(audio_url, dest)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "stock-weekly-report",
-  "version": "0.2.2",
+  "version": "0.2.4",
   "description": "Stock weekly podcast report pipeline — CLI and MCP server",
   "bin": {
     "swr": "bin/swr.js",

package/pipeline.py

@@ -253,26 +253,27 @@ def _cleanup_data_dir(data_root: Path, label: str,
         print(f"\n  Cleaned up {removed_folders} folder(s).")
 
 
-def _cleanup_audio_by_speaker(audio_root: Path, months: int) -> None:
-    """Delete audio files older than `months` from per-speaker subdirectories."""
+def _cleanup_by_speaker(data_root: Path, label: str,
+                        extensions: set[str], months: int) -> None:
+    """Delete files matching `extensions` from per-speaker subdirs older than `months`."""
     if months <= 0:
-        print(f"  Audio: retention = 0 (keep forever), skipping.")
+        print(f"  {label}: retention = 0 (keep forever), skipping.")
         return
 
-    banner(f"CLEANUP — Audio (keep {months} month{'s' if months != 1 else ''})")
+    banner(f"CLEANUP — {label} (keep {months} month{'s' if months != 1 else ''})")
 
-    if not audio_root.exists():
-        print(f"  Directory not found: {audio_root}")
+    if not data_root.exists():
+        print(f"  Directory not found: {data_root}")
         return
 
     cutoff = _cutoff_date(months)
     print(f"  Cutoff date : {cutoff}  (deleting files published before this date)")
 
     removed = 0
-    for speaker_dir in sorted(audio_root.iterdir()):
+    for speaker_dir in sorted(data_root.iterdir()):
         if not speaker_dir.is_dir():
             continue
-        for ext in SUPPORTED_AUDIO_EXTS:
+        for ext in extensions:
             for f in speaker_dir.glob(f"*{ext}"):
                 date_str = f.stem.split("_")[-1]
                 try:
@@ -281,16 +282,20 @@ def _cleanup_audio_by_speaker(audio_root: Path, months: int) -> None:
                     continue
                 if file_date < cutoff:
                     size_mb = f.stat().st_size / (1024 * 1024)
-                    print(f"  Deleting ({size_mb:.1f} MB): {speaker_dir.name}/{f.name}")
+                    print(f"  Deleting ({size_mb:.2f} MB): {speaker_dir.name}/{f.name}")
                     f.unlink()
                     removed += 1
 
     if removed == 0:
-        print("  No old audio files to remove.")
+        print("  No old files to remove.")
     else:
         print(f"  Removed {removed} file(s).")
 
 
+def _cleanup_audio_by_speaker(audio_root: Path, months: int) -> None:
+    _cleanup_by_speaker(audio_root, "Audio", SUPPORTED_AUDIO_EXTS, months)
+
+
 def cleanup_old_data(config: dict) -> bool:
     """Clean up old audio, transcript, and report files per retention config."""
     retention = config.get("retention", {})
@@ -300,7 +305,7 @@ def cleanup_old_data(config: dict) -> bool:
         parent / "audio",
         int(retention.get("audio_months", 3)),
     )
-    _cleanup_data_dir(
+    _cleanup_by_speaker(
         parent / "transcripts",
         "Transcripts",
         {".txt"},

package/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "stock-weekly-report"
-version = "0.2.2"
+version = "0.2.4"
 requires-python = ">=3.10"
 authors = [
     { name = "Chang Yu Chuan", email = "changyuchuanmicron@gmail.com" },

package/transcribe.py

@@ -100,11 +100,13 @@ def verify_transcript(transcript_path: Path) -> tuple[bool, str]:
     return True, ""
 
 
-def verify_all(audio_files: list[Path], transcript_dir: Path) -> dict[str, tuple[bool, str]]:
+def verify_all(audio_files: list[Path], transcript_root: Path) -> dict[str, tuple[bool, str]]:
     """Verify every expected transcript and return a {filename: (ok, reason)} map."""
     results = {}
     for audio_file in audio_files:
-        transcript_path = transcript_dir / f"{audio_file.stem}.txt"
+        stem = audio_file.stem
+        speaker = stem.rsplit("_", 1)[0]
+        transcript_path = transcript_root / speaker / f"{stem}.txt"
         results[audio_file.name] = verify_transcript(transcript_path)
     return results
 
@@ -170,9 +172,7 @@ def transcribe_folder(config: dict, folder_name: str) -> None:
 
     parent_folder = Path(config["parent_folder"])
     audio_root = parent_folder / "audio"
-    transcript_dir = parent_folder / "transcripts" / folder_name
-
-    transcript_dir.mkdir(parents=True, exist_ok=True)
+    transcript_root = parent_folder / "transcripts"
 
     audio_files = find_audio_files_for_run(audio_root, folder_name)
     if not audio_files:
@@ -187,7 +187,7 @@ def transcribe_folder(config: dict, folder_name: str) -> None:
     print(f"Compute type    : {compute_type}")
     print(f"Language hint   : {language}")
     print(f"Audio root      : {audio_root}")
-    print(f"Transcript dir  : {transcript_dir}")
+    print(f"Transcript root : {transcript_root}")
     print(f"Files to process: {len(audio_files)}")
     print(f"Max retries     : {MAX_RETRIES}")
     print()
@@ -199,7 +199,11 @@ def transcribe_folder(config: dict, folder_name: str) -> None:
     succeeded, skipped, failed = [], [], []
 
     for idx, audio_file in enumerate(audio_files, start=1):
-        transcript_path = transcript_dir / f"{audio_file.stem}.txt"
+        stem = audio_file.stem                          # e.g. "股癌_20260225"
+        speaker = stem.rsplit("_", 1)[0]               # e.g. "股癌"
+        speaker_transcript_dir = transcript_root / speaker
+        speaker_transcript_dir.mkdir(parents=True, exist_ok=True)
+        transcript_path = speaker_transcript_dir / f"{stem}.txt"
         label = f"[{idx}/{len(audio_files)}]"
 
         # Check if an existing transcript already passes verification
@@ -226,7 +230,7 @@ def transcribe_folder(config: dict, folder_name: str) -> None:
     # ── Post-run verification ─────────────────────────────────────────────
     print("─" * 60)
     print("Verification pass …")
-    verification = verify_all(audio_files, transcript_dir)
+    verification = verify_all(audio_files, transcript_root)
 
     all_ok = True
     for audio_name, (ok, reason) in verification.items():

package/upload_to_notebooklm.py

@@ -49,6 +49,24 @@ def default_folder_name(lookback_days: int) -> str:
     return f"{start.strftime('%Y%m%d')}-{today.strftime('%Y%m%d')}"
 
 
+def find_transcripts_for_run(transcript_root: Path, folder_name: str) -> list[Path]:
+    """Collect transcript files across per-speaker subdirs whose date falls in the run window."""
+    parts = folder_name.split("-")
+    start_str, end_str = parts[0], parts[1]
+
+    files = []
+    if not transcript_root.exists():
+        return files
+    for speaker_dir in sorted(transcript_root.iterdir()):
+        if not speaker_dir.is_dir():
+            continue
+        for f in speaker_dir.glob("*.txt"):
+            date_str = f.stem.split("_")[-1]
+            if len(date_str) == 8 and start_str <= date_str <= end_str:
+                files.append(f)
+    return sorted(files)
+
+
 # ---------------------------------------------------------------------------
 # nlm CLI wrappers
 # ---------------------------------------------------------------------------
@@ -148,19 +166,15 @@ def run(config: dict, folder_name: str) -> str:
     """Run the upload stage and return the notebook_id."""
     nlm_path = config.get("nlm_path", "nlm")
     parent_folder = Path(config["parent_folder"])
-    transcript_dir = parent_folder / "transcripts" / folder_name
+    transcript_root = parent_folder / "transcripts"
 
-    if not transcript_dir.exists():
-        print(f"ERROR: Transcript directory not found: {transcript_dir}")
+    txt_files = find_transcripts_for_run(transcript_root, folder_name)
+    if not txt_files:
+        print(f"No transcript files found in {transcript_root} for window {folder_name}")
         print("Run transcribe.py first to generate transcripts.")
         sys.exit(1)
 
-    txt_files = sorted(transcript_dir.glob("*.txt"))
-    if not txt_files:
-        print(f"No .txt files found in {transcript_dir}")
-        sys.exit(0)
-
-    print(f"Transcript folder : {transcript_dir}")
+    print(f"Transcript root   : {transcript_root}")
     print(f"Files to upload   : {len(txt_files)}")
     print()